AMBER Archive (2006)

Subject: Re: Fw: AMBER: AMBER : About the kinetic energy and temperature (GBSA model)

From: Thomas Cheatham (cheatham_at_chpc.utah.edu)
Date: Tue Mar 28 2006 - 22:32:37 CST

> > Be sure you have some experience with an "ordinary" protein or nucleic acid,
> > before embarking on a project where the initial structure might not be known
> > so well.
>
> Dear dac ,
> I set the parameter of Temperature regulation(ntt=1 and temp0=300) equal to
> "1" . It means that the temperature will be regulated around 300 K .And I
> simulated my system again .The temperature is still very high ~ 5123 k . It's
> not reasonable .Why "Sander" didn't correct the temperature?Could you give me
> more information about the temperature regulation ,such as an example ?

As DAC noted, you are strongly encouraged to gain some experience with a
more standard system, such as is explored in the series of AMBER tutorials
available on the WWW page at http://amber.scripps.edu This will allow
you to better understand the behavior you are seeing compared to what is
normal or expected.

In addition to turning on the temperature regulation, you will need to set
the coupling time (tautp=1.0). However simply controlling the temperature
will likely not overcome the large instability present in your system.
This instability is what likely leads to the very quick heating to 5000K.
You need to equilibrate the system first (minimization followed by
possibly slowly raising the temperature). If instabilities still occur,
you may have force constants too large or significant overlaps (ring
fusions) or other problems that lead to instabilities. Before you can
even diagnose these issues, you are strongly encouraged to get experience
with a smaller more standard system (like a small peptide, ACE-ALA-NME)
and encouraged to look through the AMBER archive. [Note that if you are
using GB, a Langevin temperature coupling is recommended, as is removing
center of mass rotation/translation periodically.]

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu